CONNECTING POSITIVE and Normative LEGAL THEORY

CONNECTING POSITIVE AND NoRMATIvE LEGAL THEORY

Adrian Vermeule

Positive and normative legal theory often seem to have little to do with one another. Part I describes the disconnect and suggests that it arises from two sources: the gap between fact and value, and the gap between external and internal perspectives on law. In the following Parts, I lay out a repertoire of strategies and mechanisms for connecting positive and normative legal theory. Part II examines cases in which positive theory serves as a direct source of normative arguments. Part III examines cases in which positive theory serves as an indirect constraint on normative decisionmaking. In the latter case, positive theory serves a constructive role by narrowing the set of normative arguments that must be considered when deciding what to do. Part IV extends the theme of constraints to a second-order question: In light of our best positive theories, to what audiences can normative scholarship be addressed?

I. POSITIVE AND NORMATIVE THEORY DISCONNECTED

Positive and normative legal theory (and for that matter, political and moral theory, but I will focus on legal theory) often seem radi- cally disjunct; it is sometimes not clear, even in principle, what positive and normative theory have to do with one another. Two recur- ring conceptual problems create this disconnect. The first is the gap between fact and value. The second is the gap between internal and external perspectives on law. I will offer some brief remarks on each before discussing, in the following Parts, how the gaps can be bridged. The fact-value distinction is alive and well in many quarters, despite the best efforts of some philosophical pragmatists and the dis- repute of logical positivism. Some economists remain hostile to welfare economics on positivist grounds, arguing that the Kaldor-Hicks

Professor of Law, Harvard Law School. Prepared for the University of PennsylvaniaJournal of Constitutional Law 2007 Symposium, "Positive Approaches to Constitutional Law and Theory." Thanks to Matt Stephenson, Cass Sunstein, Mark Tushnet, and the symposium participants for helpful comments, and to Abby Wood for helpful research assistance. JOURNAL OF CONSTITUTIONAL LA W [Vol. 10:2 criterion and other staples of normative welfarism require interper- sonal comparisons that are meaningless, epistemically too demanding, or unscientific. Lionel Robbins's original argument against in- terpersonal comparisons, although it conflated these conceptual, epistemic, and methodological points, argued in fact-value terms. 1 Philosophical argumentation that seeks to close or reduce the gap, such as the distinction between "brute" facts that are value free and "institutional" facts that presuppose values,2 does not fully cap- ture the everyday appeal of the distinction. Philosophers who deny that the distinction captures any metaphysical truth may also admit its utility, thus allowing that there is a pragmatic case for the fact-value distinction.3 And the distinction is part of the everyday toolkit of debates in legal theory. Thus in the debate over "super precedents, 4 originalist Randy Barnett argues in part that common-law theorists confuse fact and value: An explanation of why a particular decision will not soon be over- ruled.., is distinct from an argument for why it ought not one day be re- versed when the time is ripe .... [I] n their defense of the irreversibility of super precedents, [common-law theorists] seem to be committing two fundamental fallacies. The first is the conflation of the "is" with the "ought"; the second is the conflation of the "actual" with the "necessary."5 Barnett's first fallacy attempts to sever the connection between the positive and normative sides of precedent-based approaches to constitutional interpretation. His second fallacy addresses not the fact- value distinction, but the modal distinction between alterable facts and constraints that are inalterable (at least in the short run). I take up that distinction in Part III. The second gap is between external and internal perspectives on law. Positive theorists, particularly theorists from nonlegal disciplines or legal theorists heavily influenced by those disciplines, often take an external perspective on law, while traditional doctrinalists stick reso-

1 See generally LIONEL ROBBINS, AN ESSAY ON THE NATURE & SIGNIFICANCE OF ECONOMIC SCIENCE 136-47 (2d ed., rev. & extended 1940). 2 See generallyJohn R. Searle, How to Derive "Ought"from "Is, "73 PHIL. REv. 43 (1964). 3 See, e.g., HILARY PUTNAM, THE COLLAPSE OF THE FACT/VALUE DICHOTOMY AND OTHER ESSAYS 43-45 (2002) (critiquing various justifications for the fact-value distinction). 4 See generally Michael J. Gerhardt, Essay, Super Precedent, 90 MINN. L. REv. 1204 (2006) (arguing that some constitutional decisions have risen to the stature of "super precedents" by, among other things, taking on great importance in the public consciousness); Daniel A. Farber, Essay, The Rule of Law and the Law of Precedents, 90 MINN. L. REv. 1173 (2006) (discussing the tension between the rule of law and stare decisis). 5 Randy E. Barnett, Response, It's a Bird, It's a Plane, No, It's Super Precedent: A Response to Farberand Gerhardt,90 MINN. L. REV. 1232, 1241-42 (2006) (footnote omitted). Jan. 2008] CONNECTING POSITIVE AND NORMATIVE LEGAL THEORY 389

lutely to the internal perspective. When theorists deploy positive political models that depict "law" as an equilibrium outcome of institutional interactions, the payoff is unclear from the perspective of the old school. On this view, if a judge asks at oral argument what the "law" is in a given area, the advocate who answers, "The law is what emerges in equilibrium from your interaction with other institutions," is guilty of a category mistake. The judge is asking what she should do, and why; the external account of law is no help with that question. On this view, Holmes's famous claim that law is a predic- tion of what the courts will in fact do6 is a resolutely external account of law, one that cannot coherently be offered within legal practice. This problem can arise from the other side of the internal- external divide as well. When the internal theorist (implicitly ad- dressing the judiciary) says that the law rightly understood requires the judge to say X, but the external theorist can show with high prob- ability that the judge will say Y, or can show that a multimember court deciding a series of cases will in all likelihood say nothing coherent at all,7 it is not clear what the relevance of the internal theorist's claim really is. In Part IV, I examine the limiting case of this situation, where the normative theorist recommends public-spirited measures to officials who, according to the best positive theory, have no incen- tive or motivation to listen to such advice. These twin possibilities- that positive theory lacks a doctrinal payoff, and that normative theory lacks a connection to the real world-emphasize the conceptual gap between internal and external perspectives. These gaps between fact and value, and between internal and external perspectives, are problems, but there are solutions as well. The following Parts offer an analytic taxonomy of strategies for bridging the two gaps, with illustrations from legal theory.

II. THE NORMATIVE POWER OF THE FACTUAL

I will begin with cases in which positive theory serves as a source of normative theory; the next Part examines the constraints that positive theory places on normative theory. The cases in this Part can be lumped together under the rubric of "the normative power of the

6 See 0. W. Holmes, Justice, Supreme Judicial Court of Mass., The Path of the Law, Address at the Dedication of the New Hall of the Boston University School of Law (Jan. 8, 1897), in 10 HARV. L. REV. 457, 461 (1897) ("The prophecies of what the courts will do in fact, and nothing more pretentious, are what I mean by the law."). 7 See generally Frank H. Easterbrook, Ways of Criticizing the Court, 95 HARV. L. REv. 802 (1982). JOURNAL OF CONSTITUTIONAL LAW (Vol. 10:2 factual"8: facts or causal mechanisms are a direct source of normative arguments that would not have existed had the relevant facts or causation been different. In some cases (examined in Parts II.A and II.B), the normative power of the factual is rationally defensible. In other cases (examined in Part II.C), however, the normative power of the factual operates by nonrational mechanisms. The role of facts and causal theories in creating constraints, and thus an indirect source of normative arguments, is examined in Part III.

A. Prescriptive Theory

In common parlance, normative theory encompasses two types of arguments: claims about the best means to adopt, given stipulated ends, and claims about what ends it would be good to adopt. How- ever, claims about the best means to adopt, given stipulated ends, can usefully be given a separate label as "prescriptive" theory. Prescriptive theories support conditional arguments from is to ought. In the case of individual decisionmaking, if your goal is X, and the facts are Y, then you should do Z. In the case of social or collective decisionmaking, if there is normative consensus on X, and policy Ywould in fact produce X (ignoring other costs), then Y is a good policy. If the goal is to minimize unemployment, and the minimum wage increases unemployment, then the minimum wage is bad and should be abol- ished. Even if positive arguments about facts and causation cannot directly change people's ultimate preferences or values, surprising positive claims can supply new information that changes their derived preferences over different policies, legal rules, or institutional ar- rangements. The best positive political theory has prescriptive value of this sort. An example is Jerry Mashaw's model of the interpretive canon that constitutional questions should be avoided, if fairly possible. The canon is sometimes defended on the ground that, after all, the legislature can reinstate the constitutionally problematic statutory rule by legislating more explicitly, should a majority desire to do so. But Mashaw's treatment yields the counterintuitive implication that under a regime of narrowing interpretation, "even if the legislature acts to 'correct' an interpretation with which it disagrees, it will almost never end up with its originalpolicy reinstated, even ifnot a single member of

8 Georg Jellinek, Allgemeine Staatslehre [General Theory of the State] 338 (1929). Of course, I am ripping this out of its original theoretical context. Jan. 2008] CONNECTING POSITIVE AND NORMATIVE LEGAL THEORY 391 the legislature has altered his or her preferences."9 The main effect, one that is easy to overlook without explicit modeling, is that a narrow interpretation changes the status quo point and thus "reconfigure[s] the structure of subsequent legislative bargaining." 10 The general point is that many normative approaches are partially fact-dependent: they support no prescriptions at all unless and until facts are plugged in, although facts are only necessary, not sufficient. Famously, all purely consequentialist approaches are of this sort, because they hold that actions (or rules or dispositions) are good if and only if their factual consequences are good; a value theory specifying the good is also necessary, but facts are indispensable. Whether this is an advantage or a defect of consequentialism is an open question. Critics suggest that consequentialism is epistemically too demanding, because no consequentialist will ever have sufficient information to make the required judgments, and that consequentialism is scary, because any action, however monstrous, might be indicated by the theory, given some conceivable state of affairs. For consequentialists, however, the fact-dependence of the approach is an advantage, not least because, in many cases, it underscores that when people differ about what to do, they are often disagreeing (perhaps implicitly) about facts and consequences. Debates over the minimum wage are often driven by different estimates of its effect on employment levels and other components of the well-being of low-wage workers, rather than by high political principles.

B. Tradition as a Source of Norms

Another interpretation of "the normative power of the factual" is that tradition can serve as a source of norms. Normative theorists (including constitutional theorists) who emphasize tradition attempt to leverage from "is" to "ought." On this class of views, the way things are and have been is taken as a guide for making current decisions, perhaps because following tradition minimizes disruption and pro- tects expectations, or because it conserves decisionmaking costs, or because it aggregates the contributions of many minds, or because evolved institutions are likely to be optimal. In constitutional theory, Burkeans and common-law constitution- alists emphasize the role of traditions in giving content to constitu-

9 JERRY L. MASHAW, GREED, CHAOS, AND GOVERNANCE: USING PUBLIC CHOICE TO IMPROVE PUBLIC LAW 102 (1997). 10 Id. at 103. JOURNAL OF CONSTITFTIONAL LA W [Vol. 10:2

tional norms, understanding traditions to include both general social and political traditions, on the one hand, and narrower judicial traditions or precedents, on the other hand." Common law constitution- alism also shades into Dworkin's "law as integrity," in which precedents are part of the legal landscape into which our principled justifications must fit. 2 The language of fit, however, also makes precedents sound like a constraint on acceptable justifications rather than a direct source of norms, so integrity is an ambiguous case. The puzzles and problems surrounding traditionalism in constitutional adjudication are many.13 In the best case for traditionalists, the relevant tradition has been continuous since some point in the re- mote past, so it is currently the status quo, and the relevant tradition can uncontroversially be identified. These conditions will often fail to hold, however. For one thing, traditions notoriously can be de- scribed at higher or lower levels of generality. 4 For another, it is not obvious how Burkeans can deal with interrupted traditions. Where a tradition was followed at time 1, but was abrogated or violated at time 2, what is the good Burkean to do at time 3? Both the time 1 practice and the time 2 practice have their claims. Even if the best conditions do hold, there is a further problem that is especially troublesome."1 If judges decide on the basis of precedent, rather than using their independent judgment, a type of Burkean paradox arises. The paradox is that if many participants in the line of precedent or tradition followed the precedent or tradition (rather than exercising their independent reason), because doing so was a way to improve their information, then the informational value

11 See generally Thomas W. Merrill, Bark v. Burke, 19 HARV.J.L. & PUB. POL'Y 509, 511-12, 515-19 (1996) (identifying differences among originalism, normativism, and conventionalism); David A. Strauss, Common Law ConstitutionalInterpretation, 63 U. CHI. L. REV. 877, 884-904 (1996) (detailing a precedent-based theory of the development of constitutional law); Ernest Young, Rediscovering Conservatism: Burkean Political Theory and Constitutional Interpretation, 72 N.C. L. REV. 619, 697-712 (1994) (evaluating the case for traditionalism in constitutional interpretation). 12 See generally RONALD DWORKIN, LAW'S EMPIRE 225-75 (1986) (distinguishing "law as integrity" from other accounts, such as conventionalism and pragmatism). 13 Here I draw on the illuminating recent overview in Cass R. Sunstein, Burkean Minimalism, 105 MICH. L. REV. 353 (2006). 14 The Burkean can propose simple rules to deal with this, such as always describing the tradition at the lowest relevant level of generality. See, e.g., Michael H. v. Gerald D., 491 U.S. 110, 127 n.6 (1989) (plurality opinion). But the problem is why that simple rule, rather than another, should be chosen. 15 This paragraph and the next draw upon Adrian Vermeule, Essay, Common Law Constitu- tionalism and the Limits of Reason, 107 COLUM. L. REV. 1482 (2007). Jan. 2008] CONNECTING POSITIVE AND NORMA TIVE LEGAL THEORY 393

of the tradition is lower to that extent; there are fewer independent minds contributing to the collective wisdom. On this account, Burkean praise for precedent is self-defeating. The best contributions to the stream of precedent are those in which individual judges, or small groups of judges, exercise their unaided reason. Those who rely on tradition because it is, in Burke's words, the "bank and capital of nations and of ages," 6 make withdrawals from the common pool of information, for their private benefit; those who exercise their unaided reason contribute to the common pool, for the good of all. Despite all this, the infirmities of Burkeanism should not obscure the important theoretical role that tradition plays. Right or wrong, the claim for tradition is that the way things are and have been itself serves as a direct source of norms, potentially bridging the gap between fact and value. The appeal to tradition remains the most theo- retically critical attempt to leverage directly from "is"to "ought."

C. The NonrationalPower of the Factual

So far, we have been considering ways in which positive theory might serve as a source of norms for a rational decisionmaker. How- ever, there are also mechanisms that cause nonrational or boundedly rational decisionmakers to afford weight to the way things are- weight that is arbitrary or excessive according to some background account of rational decisionmaking. At the level of individuals, the umbrella label is status quo bias, which subsumes phenomena such as loss aversion, under which losses from an arbitrary baseline are weighted more heavily (in fact about twice as heavily) as equivalent gains;17 the related phenomenon of opportunity-cost neglect, in which people weigh out-of-pocket costs more heavily than foregone gains, even if the two are economically equivalent; and the endowment effect, which under certain circumstances may cause a diver- gence between individuals' valuations of things they possess and 8 things they could obtain.1

16 EDMUND BURKE, REFLECTIONS ON THE REVOLUTION IN FRANCE 74 (Frank M. Turner ed., Yale Univ. Press 2003) (1790). 17 See generally Amos Tversky & Daniel Kahneman, Loss Aversion in Riskless Choice: A Reference- Dependent Model, 106 Q.J. ECON. 1039 (1991) (describing prospect theory). 18 It is not clear that the endowment effect exists; at a minimum, it appears to be highly sen- sitive to the precise details of the situation. For a recent skeptical treatment, with an overview of the literature, see generally Charles R. Plott & Kathryn Zeiler, The Willingness to Pay-Willingness to Accept Gap, the "Endowment Effect, "Subject Misconceptions, and Experimen- tal Proceduresfor Eliciting Valuations, 95 AM. ECON. REV. 530 (2005). JOURNAL OF CONSTIFUTIONAL LAW (Vol. 10:2

Confusion arises when skeptics point out that, in certain cases, a fully rational individual may display similar behaviors, because (for example) the individual is risk averse. The problem is that the actual mechanisms that bring about the behavior may, in particular cases, have nothing to do with such post hoc redescriptions. Some behavior is overdetermined, in the sense that either the fully rational or the boundedly rational decisionmaker may engage in it, for different reasons; it is unclear why we should care that a fully rational decisionmaker might have done the same thing, if in fact the behavior did not arise in a fully rational way. The behaviors might be observationally equivalent taken in isolation, but careful experimentation can often trace the processes that produced them and determine whether those processes were or were not rational.

III. POSITIVE THEORY AS A CONSTRAINT

Besides serving as a direct source of norms, facts and causal mechanisms can serve as indirect constraints on decisionmaking. Of course there are other indirect constraints; among these are the rules of logic, as exemplified by Arrow's Theorem, 9 and other impossibility results. However, I will restrict myself to the role of positive constraints. Roughly, facts and causal mechanisms can constrain normative decisionmaking either by raising the costs of given actions, or by putting them outside the opportunity set altogether. These two ways can be reduced into one by defining actions outside the opportunity set as having an infinite cost, but this is a semantic reconciliation only, so I will treat costs and constraints separately. First, the existence of a status quo point may create costs of transition from the status quo to a preferred state; in some cases, those costs may overbalance the gains from the switch. Transition costs provide a rational interpretation of the theorist's common gambit of attempting to "shift the burden" to the other side. This is often offered in a rhetorical or tendentious spirit. In some cases, however, the other costs and benefits of a normative proposal are uncertain, and the existence of transition costs serves as a tiebreaker that deter- mines the argument in favor of the status quo. This is at the level of legal theory; within legal practice, as inside a courtroom, the shifting of burdens is straightforwardly rational, a device to allocate the costs of decisionmaking and of error.

19 See generally KENNETHJ. ARROW, SOCIAL CHOICE AND INDIVIDUALVALUES 46-60 (1951). Jan. 2008] CONNECTING POSITIVE AND NORMATIVE LEGAL THEORY 395

Second, and more strongly, facts and causal mechanisms may limit the opportunity set (or choice set or feasible set) of constitutional rules or policies." One way in which facts and causal mechanisms constrain normative theory is through the maxim "ought implies can." It is a hoary proposition of individual-level ethics that one has no obligation to do what one cannot do. It is not obvious that this maxim is correct; consider that it is fully coherent to say things like, "I should have helped him, but I couldn't." Statements of that sort suggest a sharp distinction between our first-best obligations and our second-best opportunities. On this view, although the existence of constraints might excuse compliance with our obligations, it does not eliminate their existence. 21 Pragmatically, however, the consequences for decisionmaking are the same on either view. There is no practical sense in urging a course of action that cannot be carried out. In deploying arguments that attempt to dismiss some normative proposal as infeasible or (more strongly) utopian, several further distinctions are necessary. One is between the short run and the long run. Constraints that are inelastic in the short run often turn out to be elastic in the long run, so the appeal to constraints is often invalid if offered as an objection to the first steps in a long-term campaign for change. Another distinction is between hard constraints of tech- nology, biology, and physics-no human can be ten feet tall-and the softer constraints of politics. The latter typically arise from interaction equilibria among political actors and are highly susceptible to disruption by sudden exogenous shocks or endogenous change, as when threshold effects and tipping points create sudden discontinui- ties in institutional behavior. In politics, change is often dismissed as unthinkable, up until the moment at which, suddenly, it occurs. Ar- guments from infeasibility that overlook these distinctions are a type of political fallacy2 that are often invoked to prop up the status quo for disreputable reasons.

20 Here I draw on the illuminating treatment in Lawrence B. Solum, ConstitutionalPossibili- ties (Ill. Pub. Law & Legal Theory Research Papers Series, Research Paper No. 06-15, 2007), availableat http://ssrn.com/abstract=949052. 21 "[T] he fact that an agent cannot be blamed for doing A does not show that no wrong was committed, and no norm violated." Robert Stem, Does 'Ought'Imply 'Can'? And Did Kant Think It Does , 16 UTILITAs 42, 47 (2004). 22 See generallyJEREMY BENTHAM, BENTHAM'S HANDBOOK OF POLITICAL FALLACIES (Harold A. Larrabee ed.,Johns Hopkins Press 1952) (1824) (describing politicians' use of the fallacies of authority, danger, delay, and confusion to perpetuate the status quo). JOURNAL OF CONSTITTIIONAL LA W [Vol. 10:2

IV. THE AUDIENCE FOR NORMATIVE THEORY

Finally, we may ascend to a higher theoretical level and examine the role of positive theory as a constraint on the appropriate audiences for normative theorizing.3 I have in mind various versions of what is called, in economics, the "determinacy paradox." If government is understood as a benevolent maximizer of social welfare, the theorist's public-spirited or welfare-maximizing proposals are addressed to the right audience. Suppose, however, that governmental motives are endogenized, and that government officials are modeled as rationally self-interested actors. Then it is not clear that anyone will be listening to public-spirited proposals; the audience to whom they are addressed will be motivated to adopt them only if they hap- pen to correspond to officials' self-interested aims. "[I]f what gov- ernments do is the result somehow of equilibrium behavior of self- interested actors, then advising government is as senseless an activity as -advising monopolists 24 to lower prices or advising the San Andreas fault to be quiet., The determinacy paradox does not rule out all forms of normative theorizing. Prescriptive theory is still possible; the theorist can advise self-interested actors about how to pursue their interests, and may be able to persuade them that a proximate goal they previously favored does not serve their ultimate or long-term interests. But normative advice about the ultimate ends to be pursued will either fall on deaf ears, if it diverges from the ends the self-interested agents already hold, or will be otiose, if it corresponds to those ends. In this standard version, the determinacy paradox arises because of the theorist's motivational assumptions about the decisionmakers to whom the normative proposals are addressed. However, a similar paradox arises where normative proposals are addressed to decisionmakers who are assumed to be subject to cognitive limitations. In a simplistic example, the advice not to panic during emergencies is not very useful, if addressed to decisionmakers who are assumed to have a propensity to panic during emergencies. Even knowing their own tendencies in the abstract, the decisionmakers may reason, "It's generally good advice not to panic, but this situation really is serious, so

23 This paragraph draws on a discussion in Adrian Vermeule, Self-Defeating Proposals: Acker- man on Emergency Powers, 75 FORDHAM L. REv. 631, 637 (2006). 24 Brendan O'Flaherty & Jagdish Bhagwati, Will Free Trade with PoliticalScience Put Normative Economists Out of Work, 9 ECON. & POL. 207, 207 (1997). Jan. 2008] CONNECTING POSITIVE AND NORMATIVE LEGAL THEORY 397 we must take severe measures"-a conclusion that may just embody the very same panicky tendencies that the theorist diagnosed. The determinacy paradox is ubiquitous in legal theory, particularly in public choice theory, which endogenizes the motivations of government officials at the risk of rendering public-interested proposals fruitless. With modifications, the problem applies to normative proposals in traditional doctrinal scholarship addressed (as is typically the case) to the Justices of the Supreme Court. Proposals in this vein typically attempt to bring normative coherence to some do- main of law through the Dworkinian procedure of fit and justifica- tion-reconciling conflicting cases, texts, and other legal sources, and by putting them in a normatively attractive light. In public choice scholarship, and in the attitudinal models of judicial motivations prevalent in political science,25 however, Justices 6 are assumed to engage in self-interested behavior, or at best, to vote in accordance with their notions of good public policy, constrained at most by strategic calculations of the reactions of other institutions. If this is the case, it is not obvious that anyone on the Court is listening to the doctrinal theorist's proposals for advancing the normative coherence of constitutional law. To make sense of doctrinal scholarship, one must assume either that the Justices are directly interested in creating good law, or else that the Justices are indirectly interested in doing so, because they care about their reputation with (among others) the law professors who care about good law.27 In the latter case, the Justices are a real audience for doctrinal scholarship because law professors are a real audience for the Justices' opinions. Whether this happy equilibrium exists is an open question. If it does, however, it provides a bridge be-

25 I am oversimplifying slightly. A judge, for example, might act strategically in order to maximize the chances that his (sincerely held) view of what the law requires will actually be adopted by the court and will not be overridden by other institutions. More com- monly, political scientists model judges as both strategic and interested solely in advancing their views of good policy, as opposed to good constitutional law. See, e.g.,LEE EPSTEIN &JACK KNIGHT, THE CHOICESJUSTICES MAKE 22-107 (1998) (describing the strategic activities Justices undertake to advance their preferred policy positions). 26 This assumption does not necessarily apply to lower court judges, who face many fewer cases in which the legal materials are indeterminate, and who thus have less scope for self-interested or ideological decisionmaking. 27 See LAWRENCE BAUM, JUDGES AND THEIR AUDIENCES: A PERSPECTIVE ON JUDICIAL BEHAVIOR 100-02 (2006) ("UJ]udges... care about the opinions of legal academ- ics.... [Liaw professors control sources of judicial satisfaction such as publication of opinions in casebooks and opportunities to lecture at law schools."). In Baum's picture, law professors are one of the audiences that reputation-minded judges care about im- pressing or pleasing. 398 JOURNAL OF CONS77TUTIONAL LAW [Vol. 10:2 tween internal legal theory offered by law professors and the external perspective on judicial motivations-even assuming, from an external perspective, that the Justices are not directly motivated to make the law better.